Bill Parker over at Chinleana has covered this here, but I figured I’d highlight somethin he says, and something the paper notes.
The new archosauromorph Azendohsaurus madagaskariensis  supplements our previosu taxon Azendohsaurus laaroussi [2,3], and provides damning evidence against the concept of Azendohsaurus being either an ornithischian  or a sauropodomorphan  dinosaur. While an abstract and presentation in 2002  initially disputed its dinosaurian affinities, this hasn’t been published, although it intended to do so on the basis of Moroccan remains of the original species. This taxon derives from the Triassic of Madagascar, and expands the range of the taxon Azhendousaurus.
First, Bill writes:
If this discovery does not finally demonstrate the peril of assigning isolated jaw fragments and teeth to various dinosaurian subgroups, I do not know what will. The placement of Azendohsaurus as a basal archosauromorph demonstrates that herbivory has evolved independently numerous times within Archosauromorpha and was actually much more common in this clade than previously believed. Some of the primitive cranial features found in Azendohsaurus include a pineal opening, an incomplete lower temporal bar, and palatal teeth. One unique feature of Azendohsaurus is that the palatal teeth are actually leaf-shaped with denticles, very similar to the marginal teeth.
There are a lot of morphological hints at the development of herbivory, and one of them is the development of large, triangular denticles on the crowns (the size ratio to crowns has an explicit value, but the term “coarse” has been used for this degree of denticulation, although it is — ahem — coarsely applied).
“Lanceollate” is also a term that has been given to various teeth and has been linked to herbivory, but givewn that it has a broad and inconsistent association with phylogeny and with herbivorous diet, one must really wonder about how “lanceollate” crowns are really indicative of much. No one (and I mean no one) has performed an analysis linking function to the form; it’s been assumed largely from the get go.
But let’s take distribution here: In dinosaurs alone, “lanceollate” crowns are known in “prosauropod” sauropodomorphan dinosaurs (a gradational concept of non-sauropod sauropodomorphans), “hypsilophodontan” ornithischian dinosaurs (another gradational concept of non-ornithopodan ornithischians), and even some theropod dinosaurs such as therizinosauroids. Its known in some near-dinosaurian dinosauromorphans, such as Silesaurus opolensis, but absent for most others. It may even represent the basal condition among dinosaurs, present in its outgroup and its basal theropods, sauropodomorphans and ornithischians .
Thus the finding of these teeth should be less than exceptional to recovery of a particular group of archosaurs.
Secondly, I am particularly intrigued by a comparison  make to another Madagascar taxon, Archaeodontosaurus descouensi , originally styled as a “prosauropod” sauropodomorphan, but with particularly similar dentition; both taxa were recovered from nearly the same horizon and from nearly the same place. Now, I don’t think they are the same, and I don’t think Flynn et al.  think this. It’s a handy comparison, because it indicates morphological similarity in what are likely disparate taxa in the same place. There are reasons to doubt the association systematically, largely due to the implantation in the teeth: Basal archosauromorphans like Azendohsaurus exhibit what is called ankylosed thecodonty, in which the periodontal ligament is ossified and roots the teeth into their sockets, while all dinosaurs exhibit true thecodonty (the ligamental attachment is only at the base of the root, deep in the socket. In addition, there remains a distinct lingual sheet of bone with foramina forming the lingual dentigerous wall which, in saurischian dinosaurs, is separated into individual interdental plates and loose from the remaining bone forming the dental sockets. Buffetaut  indicates that Archaeodontosaurus possesses true thecodonty, while at the same time appears to be ambiguous with regards to the presence of a whole or partite lingual dentigerous wall. These taxa may be very similar to one another, but at least one thing is clear:
Dinosaurian taxa based on teeth (and especially “common” morphologies like “ziphodont” and “lanceollate” teeth) are less likely to be securely establishable on that reason alone. Revueltosaurus, another taxon based solely on teeth  has been revealed to be a particularly non-dinosaurian crurotarsan , although the original Azendohsaurus laaroussi  is based on a partial mandible rather than just teeth. This results in another scrappily preserved taxon to be found to be nondinosaurian on the basis of more substantive reasoning than the causal morphology of its teeth, and that is a very specific warning.
 Flynn, J. J., Nesbitt, S. J., Parrish, J. M., Ranivoharimanana, L. & Wyss, A. R. 2010. A new species of Azendohsaurus (Diapsida: Archosauromorpha) from the Triassic Isalo Group of southwestern Madagascar: cranium and mandible. Palaeontology 53:669-688.
 Dutuit, J. M. 1972. Decouverte d’un dinosaure ornithischien dans le Trias superieur de l’Atlas occidental marocain [Remains of an ornithischian dinosaur from the Upper Triassic of the western Atlas of Morocco]. Comptes Rendus d l’Academie des Sciences, Paris 275:2841–2844.
 Gauffre, F. 1993. The prosauropod dinosaur Azendohsaurus laaroussii from the Upper Triassic of Morocco. Palaeontology 36:897–908.
 Jalil, N.-E. & Knoll, F. 2002. Is Azendohsaurus laaroussii (Carnian, Morocco) a dinosaur? Journal of Vertebrate Paleontology 22(suppl. to no. 3):70A.
 Sereno, P. C., Forster, C. A., Rogers, R. R. & Monetta, A. M.1993. Primitive dinosaur skeleton from Argentina and the early evolution of Dinosauria. Nature 361:64-66.
 Buffetaut, E. 2005. A new sauropod dinosaur with prosauropod-like teeth from the Middle Jurassic of Madagascar. Bulletin de la Société Géologique de France 176:467–473.
 Hunt, A. P. 1989. A new ?ornithischian dinosaur from the Bull Canyon Formation (Upper Triassic) of east-central New Mexico. p. 355-358 in Lucas and Hunt (eds.) Dawn of the Age of Dinosaurs in the American Southwest. (New Mexico Museum of Natural History, Albuquerque.)
 Parker, W. G., Irmis, R. B., Nesbitt, S. J., Martz, J. W. & Browne, L. S. 2005. The Late Triassic pseudosuchian Revueltosaurus callenderi and its implications for the diversity of early ornithischian dinosaurs. Proceedings of the Royal Society B 272:963-969.
Buffetaut described Archaeodontosaurus as Middle Jurassic, which makes it much more recent than Azendohsaurus. Has anyone disagreed with this?
The material comes from the Isalo IIIb beds. I am not versed on the Isalo stratigraphy, but my assumption is that the stratigraphy is accurate. Flynn’s 1999 paper in Science implies this is accurate.
Archaeodontosaurus is from the Isalo III beds, but Azendohsaurus is from Isalo II. I guess it is possible that Azendohsaurus-like reptiles survived into Isalo III, but I wouldn’t call a difference of tens of millions of years “nearly the same horizon.”
Your blog’s coming along quite well. Interesting thought on Archaeodontosaurus. You state that “Dinosaurian taxa based on teeth (and especially “common” morphologies like “ziphodont” and “lanceollate” teeth) are less likely to be securely establishable…”, but both Azendohsaurus and Revueltosaurus turned out to still be valid taxa, even though they were misidentified originally.
Revueltosaurus callenderi was part of my original list of dinosaurian taxa based on teeth that would effectively be rendered nomina dubia. I know Bill Parker would also disagree with this. The teeth seem to be distinct … to a point. With description of the complete skull of the crurotarsan, teeth referred to Pekinosaurus, Galtonia and Revueltosaurus were found to correspond to various sections of this taxon’s jaw. The decision was made to refer the skeletal material to the dental taxon, a decision I wish to approach in more detail (and must) when it comes time to fully detail my approach to TBT.
But you’re now arguing that in hindsight, the skull having these teeth should validate the nomenclature basedf on the teeth. What if Revueltosaurus hunti is found to have a skull identical to that of Revueltosaurus callenderi, or that another revueltosaur will have teeth that are different, but overlap with the types of either callenderi or hunti? Not only is this possible, it occurs in sharks. We are fortunate that most shark dentitions come found in clusters of finds, but this isn’t always the case, nor was it the case with most conodont assemblages. We assume the teeth as we find it form a series (as Thulborn did with Phyllodon) regardless of any actual association. What’s interesting is that when Rogers et al.  described cranial material of what is now called Majungasaurus crenatissimus, they referred it to Majungatholus atopus because Majungasaurus crenatissimus is based on loose teeth (to which a jaw is referred, to which the rest of the crania is referred as a result). They later flipped on this identification, because of the referred mandible.
Very tricky stuff, and hardly secure when the types of these taxa are teeth which are asserted after the referral of more extensive material.
 Rogers, R. R.; Krause, D. W. & Curry-Rogers, K. 2007. Cannibalism in the Madagascan dinosaur Majungatholus atopus. Nature 422:515–518.